Watch the Navy’s Robotic Gunboats Swarm the James River
Thirteen robotic boats simulate attack
The Navy has long wanted to get into the robot game. The admirals are envious of the Air Force’s hundreds of flying drones, and the sailing branch is working on a carrier-launched drone of its own.
But now the Navy is creating swarms of robotic gunboats and testing them on the James River in Virginia.
On Aug. 14, the Office of Naval Research unleashed 13 robotic boats on a mission to intercept a simulated enemy vessel. Once the researchers gave the order, the machines planned and carried out an attack pattern by themselves.
The machines closed in on the enemy ship and began circling it. The only thing the Navy didn’t do was ordering the robots to open fire.
“They can overwhelm that threat, and frankly, they can destroy that threat if necessary,” ONR chief Rear Adm. Matthew Klunder said. “And frankly, that’s what this technology is about.”
It wasn’t the first time the Navy experimented with unmanned surface ships. But it was the first time it fielded so many at once. The Navy also claims it has the working technology to command a networked swarm of more than 30 robotic boats at the same time.
The vessels—armed with a mix of 50-caliber machine guns and high-powered microwaves—were largely autonomous. They can follow and intercept both static and moving vessels. One of the few things the machines can’t do on their own is make the decision to shoot.
“We have every intention of using those unmanned systems to enable our sailors to engage a threat, and destroy it if necessary,” Klunder said. “But let me anchor on one really important factor here that’s critical to us. There is always a human in the loop of designation of a target, and if so be it, destruction of the target.”
The trick to making it all work is what the Navy calls CARACaS, or Control Architecture for Robotic Agent Command and Sensing. The mix of sensors—both electro-optical and infrared—and CARACaS’s networked computer brain allow a single sailor aboard a warship to give commands to the entire swarm.
Target an enemy ship, and the swarming robots take off on their own, using algorithms to determine the optimal attack pattern. But the human controller can take over for more precise instructions.
If it works out, Navy vessels transiting congested harbors and dangerous straits could soon deploy swarms to ward off waterborne suicide attackers or waves of enemy speedboats.
The Navy is also betting on the machines being cheap. It only costs a few thousand dollars to convert an existing patrol boat into a CARACaS-powered robot. The Navy has hundreds of these patrol boats sitting in its inventories.
To be sure, this doesn’t matter that much against the big threats facing the Navy in the foreseeable future—such as the Chinese land-based anti-ship missiles and submarines increasingly proliferating around the world.
Instead, the Navy’s push into unmanned ships is aimed at small threats, like the suicide attack on the destroyer USS Cole. As the Cole rested in a Yemeni harbor, terrorists crept up in a small boat and blew a hole in the ship, killing 17 sailors and nearly sinking the vessel. In theory, a swarm of robot boats could have intercepted the terrorists before they came close enough.
Swarming robot gunboats could also act as a counter against large numbers of human-driven, enemy gunboats.
The Iranian navy trains for this kind of asymmetric warfare at sea. Tehran can’t possibly rival the U.S. on the water. But a rush of cheap speedboats full of sailors armed with rocket-propelled grenades—and anti-ship missiles—can be a serious threat to a large and expensive American warship designed to fight other large and expensive warships.
Iran is also adding swarms of light airplanes and helicopters to the mix. It would not be enough to defeat the United States—far from it. But it might be enough to give Washington a bloody nose.
The robotic boats allow the Navy to come close to meeting Iran’s surface threat on an equivalent basis. At the same time, CARACaS allows as few as a single sailor to control the entire swarm, freeing up sailors who would normally be piloting the patrol boats to handle other tasks.
But that poses something of a dilemma.
Note the Navy is experimenting with armed robots. Devolving decision-making for as many as 30 robots to a single person is fraught with danger. That’s even more the case in crowded harbors and straits where potential attackers lurk among civilian vessels.
The robots can’t make the decision to fire on their own. That’s the job of the human controller. But how can one person decide which of a dozen, two dozen or more boats should be the ones to fire—while avoiding hitting neutral civilian vessels in the same area?
The Navy insists it has safety protocols built into CARACaS to deal with these unexpected problems. The other idea is to not use huge numbers of boats all the time. If there’s a threat, “This doesn’t mean I send all 20 over there if that’s not what is needed,” Klunder said. “It may be that you only need to send five or six.”